HAF uses an LLM agent and deadline-aware convex allocation to reach 90% SLO fulfillment in AI-RAN, improving AI request fulfillment from 51% to 85.3%.
Yinyangran: Resource multiplexing in gpu -accelerated virtualized rans
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DC 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Deadline-Driven Hierarchical Agentic Resource Sharing for AI Services and RAN Functions in AI-RAN
HAF uses an LLM agent and deadline-aware convex allocation to reach 90% SLO fulfillment in AI-RAN, improving AI request fulfillment from 51% to 85.3%.